Man I love the cooling system sharky lol do what you gotta do right. All part of R&D. Glad to hear your making even more progress on these things. Paving the way for the rest of us man. Keep up the good work brother!!!
Check out W6QPL website, I design these for high reliability , typical 5 KW Pout , the trick is use ferrite that does not saturate. Saturated ferrite will get hotter temperature. Higher Output Power means lower impedance , meaning use lower lower impedance coax, But Higher output power can Arc, so use bigger Ferrite cores and coax.
You showed the harmonics on the scope. I see no LPF. Have you done the same on the scope with and without an LPF? I am now a disabled Vet, coming from electronic engineering and tech career, will set up some equipment at home to play with LDMOS stuff for fun and experimental purposes. Thanks. Was looking at the same scope and generator you have at Amazon.
BTW, you're right... heat sink paste for those very high power devices is inadequate. You need to solder those devices to a copper heat spreader. Then, the spreader mounts to the sink with paste. That combination has a much larger surface/contact area.
Having spent the last few months learning everything I can about thermal solutions, and along with many hours of CFD simulations, I've come to the conclusion that paste (good paste, 4w-mk or better) is acceptable if TDP is around 500w or less. Beyond that... soldering along with a highly engineered cooling system are mandatory.
@@MudDuckSharky I don't think it's the 500W that's the barrier for the paste... it's the power density. If those very high power devices had the internal die spread out... and were twice as big at the flange... then you could get away with the paste.
@@MudDuckSharky I think that wet sanding both surfaces, transistors and copper, is the first step. Add a thin film of thermal silicone and you are done. Regarding reliability, you must add a fast switching removing driving power whenever a fast peak of over driving power appears, it looked to me that your radio sends a power peak before selected driving power settles down to get at the output those 2100 PEP. I have four power mosfet that remove power supply for each device if there is a spike of excessive driving power high SWR, more than 35A at 50 volts, or high temperature. A big surface of fresh air at the front panel and two 170mm 48 blowers, runing very slow at only 13 volts, extracting air at the rear keep temperature of LDMOS below 55°C. Fresh air comes from the front cooling the top of the amplifier and the heatsink below as well. Heatsink is huge, amplifier weight is 40Kg not including the external power supply. Regarding linearity I have set my idling current to 2.5A for each of four my mrfe6vp61k25h devices, after having seen the imd plots of these devices. That's 125 watts each, 500w in all, of heat before any rf power is applied but I get decent linearity limiting the output power to 500w. My 3 to 50MHz linear amplifier is built using four modules, four 1250w devices, to get 2Kw output quite clean with imd3 35dB below the levels of each of the two test audio tones. 4Kw is possible but I must keep power below 2.5Kw for linearity to be good enough for not splattering 9KHz when my USB audio bandwidth is only 3KHz. 2500mA for each device may seem a lot, but if you look at the plots of the Freescale data sheet you'll see that 4500mA gives best linearity if output power is limited low enough. BTW... I haven't smoked a single LDMOS... Not yet!
Since a harmonic cannot be lower than the fundamental frequency, I think you are looking at sidebands of the modulated signal as well as harmonics... Your results are actually better than you thought they were! #12
